RE: linux-next: Tree for Aug 17 (staging/hfi1)

[Marciniszyn, Mike]

> A:  these and other similar errors:
> 
> In file included from ../drivers/staging/hfi1/chip.c:61:0:
> ../drivers/staging/hfi1/chip.c: In function ÃÂÂ__hfi1_trace_LINKVERBÃÂÂ:
> ../drivers/staging/hfi1/trace.h:1357:20: error: function
> ÃÂÂ__hfi1_trace_LINKVERBÃÂÂ can never be inlined because it uses variable
> argument lists  static inline void __hfi1_trace_##fn(const char *func, char
> *fmt, ...) \
>

This was corrected with
https://www.mail-archive.com/linux-rdma@xxxxxxxxxxxxxxx/msg27221.html, which I think Doug has picked up.

Doug, Randy, Stephen how do we want to handle fixing drivers/staging/hfi1 via the staging list vs. the linux-rdma list?

> B:  When CONFIG_PCI is not enabled, lots of these errors:
> 

The driver used to get a gift from the drivers/infiniband/hw/Kconfig, which has the required PCI dependency.

The infiniband dependency is also similarly missing and was resolved via the following temporary patch:

https://www.mail-archive.com/linux-rdma@xxxxxxxxxxxxxxx/msg26989.html

This one also needs to be similarly added for that 0-day build issue.

I have submitted to both lists https://www.mail-archive.com/linux-rdma@xxxxxxxxxxxxxxx/msg27299.html assuming the above one as a dependency.

> In file included from ../drivers/staging/hfi1/chip.c:60:0:
> ../drivers/staging/hfi1/hfi.h:508:20: error: field ÃÂÂmsixÃÂÂ has incomplete
> type
>   struct msix_entry msix;
>                    ^
> 
> ../drivers/staging/hfi1/chip.c: In function ÃÂÂdisable_intxÃÂÂ:
> ../drivers/staging/hfi1/chip.c:8667:2: error: implicit declaration of function
> ÃÂÂpci_intxÃÂÂ [-Werror=implicit-function-declaration]
>   pci_intx(pdev, 0);
>   ^
> 
> so the driver should depend on PCI or maybe even PCI_MSI.
> 
> C.  Please provide some contact info, e.g. in a TODO file or the MAINTAINERS
> file.
> 

I have submitted https://www.mail-archive.com/linux-rdma@xxxxxxxxxxxxxxx/msg27298.html to both lists to add a maintainer.

Mike